J. Beauté, CEREDIH (Centre de Référence Déficits Immunitaires Héréditaires), Groupe Hospitalier Necker-Enfants Malades, 149 rue de Sèvres, 75743 PARIS Cedex 15, France. E-mail: email@example.com
Lifelong immunoglobulin replacement is the standard, expensive therapy for severe primary antibody deficiencies. This treatment can be administrated either by intravenous immunoglobulin (IVIG) or subcutaneous infusions (SCIG) and delivered at home or in an out-patient setting. This study aims to determine whether SCIG is cost-effective compared with IVIG from a French social insurance perspective. Because both methods of administration provide similar efficacies, a cost-minimization analysis was performed. First, costs were calculated through a simulation testing different hypothesis on costs drivers. Secondly, costs were estimated on the basis of field data collected by a questionnaire completed by a population of patients suffering from agammaglobulinaemia and hyper-immunoglobulin (Ig)M syndrome. Patients' satisfaction was also documented. Results of the simulation showed that direct medical costs ranged from €19 484 for home-based IVIG to €25 583 for hospital-based IVIG, with home-based SCIG in between at €24 952 per year. Estimations made from field data were found to be different, with significantly higher costs for IVIG. This result was explained mainly by a higher immunoglobulin mean dose prescribed for IVIG. While the theoretical model showed very little difference between SCIG and hospital-based IVIG costs, SCIG appears to be 25% less expensive with field data because of lower doses used in SCIG patients. The reality of the dose difference between both routes of administration needs to be confirmed by further and more specific studies.
Antibody deficiencies are the most frequent causes of primary immunodeficiency diseases (PID), accounting for approximately 60–70% of all PID . These deficiencies are a heterogeneous group of diseases which are usually classified into four main syndromes: agammaglobulinaemia, hyper-IgM syndromes (HIGM), hypogammaglobulinaemia including common variable immunodeficiency (CIVD) and selective immunoglobulin deficiencies [2,3]. These diseases share an increased susceptibility to potentially life-threatening infections, although the extent of this predisposition is variable, depending upon the depth of the deficiency. Lifelong immunoglobulin replacement therapy has been shown to be efficient and therefore has become the standard treatment for more than four decades . Immunoglobulin replacement therapy can be administrated either by intravenous immunoglobulin infusions (IVIG) or by subcutaneous infusions (SCIG). Patients usually receive IVIG in an out-patient setting with an infusion every 21 or 28 days. In France, although possible, IVIG is seldom home-delivered. Conversely, SCIG is given at home in most cases with an infusion every 7 days. Because immunoglobulin replacement is a lifelong and expensive medication, it has to be assessed from an economic perspective. Whether one route of administration could be more efficient than another needs to be determined. Previous studies suggest that IVIG and SCIG offer equivalent protection against infections and that adverse events are negligible [5–7]. However, the small size of the populations compared and methodological insufficiencies have been underlined recently by Canadian authors . In terms of quality of life, there is evidence showing that patients with agammaglobulinaemia receiving immunoglobulin replacement have a quality of life close to the general population . Some results suggest that patients with home-based SCIG perceive an improved quality of life [10,11]. Regardless of the method of administration, costs are driven mainly by the cost of the Ig preparation that accounts for nearly 80% of total costs . However, results of previous economic studies carried out in different countries (France, Germany and Sweden) have drawn diverse conclusions that may reflect the specificities of the local health systems [7,8,12,13]. Hitherto, only one study has been performed in France , and has not been considered to be an economic evaluation but merely a cost study by the Centre for Reviews and Dissemination of the University of York. This study aims to determine whether SCIG at home is cost-effective compared with IVIG delivery at hospital from a French social insurance perspective. The patient's perspective was also considered, including assessment of out-of-pocket payments and satisfaction.
Based upon a systematic literature review, the assumption was made that both treatments provide identical clinical efficacy. If Ig replacement gives patients a normal quality of life, it can be assumed that quality adjusted life years (QALYS) of treated patients are close to 1. Hence, a cost minimization analysis was chosen. For out-patient treatment, hospital costs and costs of transportation were taken into account. For home care, cost of nurse care and cost of additional devices such as infusion pumps were also considered. The baseline case is a 50 kg young adult living 15 km from the hospital with a nurse 15 km from his home. The infusion dose for IVIG is 0·45 mg/kg every 21 days and 0·15 mg/kg every 7 days for SCIG. It is assumed that both methods of infusion require equal monthly doses. The cost of immunoglobulin is fixed by the French authorities at €45/g for both treatments. In France, pumps are rented and French social insurance has fixed the amount of reimbursement at €306·41/month per unit. Hospital costs do not include check-up visits, because the number of these visits is similar for all modalities (one visit every 6 months). Home therapy requires training in hospital. In France, time of training is not well defined and may differ from one hospital to another. Our experience suggests an average time of training of 4 h provided by a nurse. This represents an amount of €110. Because this spending is low and occurs only once, it was not taken into account. As for any patient with severe PID, comprehensive coverage is offered by social insurance. Costs were taken as provided by French social insurance. The cost of time lost was calculated upon a base of a French mean gross salary provided by the French National Institute for Statistics and Economic Studies (INSEE). The amount was €2600 per month for 35 h full-time work in 2006. The gross salary was chosen according to the human capital approach .We assumed that half a day was necessary for out-patient treatment, while 1 h was lost for home-based infusion. A time horizon of 1 year was chosen. To test the robustness of the results, several one-way sensitivity analyses were carried out.
To test our model with field data, a questionnaire was given to a cohort of patients with congenital agammaglobulinaemia (Bruton's disease or autosomal recessive agammaglobulinaemia) or hyper-IgM syndrome. These two diseases share similar clinical patterns with absence of IgG production. The cohort included patients followed by physicians who participate in the French National Centre for Primary Immunodeficiencies (CEREDIH). All patients having an IgG level below 2 g/l at the time of diagnosis were asked for participation and enrolled into the study. No age limitation was set, as parents were asked to complete the questionnaire on the behalf of their child when necessary. Data collection started in February 2008 and lasted until October 2008. The questionnaire comprised a short set of questions about the modalities of the treatment, including data on direct medical cost and data on indirect cost. Out-of-payments were taken into account. The assumption made here was that even with theoretical full coverage, some burden may still rely on patients. For instance, one patient may be bound to take a day off to get his medication or to plan his journey to the hospital. Patients' treatment satisfaction was addressed using the Treatment Satisfaction Questionnaire for Medication (TSQM) version II for its simplicity and previous validation [15,16].
It is assumed that people prefer to spend their money and enjoy its benefits now rather than in the future. Therefore, discount rates are usually applied to both costs and outcomes [17,18]. Because there was merely one period in our treatment (Ig replacement is assumed to be constant over time with a continuous effect), no discount was calculated.
Continuous variables were given with their median and interquartile range (IQR) or with their mean. Variables were compared across groups using the Mann–Whitney U-test or Student's t-test for continuous variables, and the χ2 or Fisher's exact test for categorical variables.
For all analyses, statistical significance was defined as P < 0·05. Statistical analyses were performed using sas version 8·02® (SAS Institute Inc., Cary, NC, USA).
Direct medical costs ranged from €19 484 for home-based IVIG up to €25 583 for hospital-based IVIG, with home-based SCIG in between at €24 952 per year (Table 1). Thus, €797/year could be gained in theory by switching from IVIG hospital-based to SCIG and €6099 from hospital-based to home-based IVIG. Parameters used for one-way sensitivity analyses are displayed in Table 2. Periodicity of immunoglobulin replacement, need for nurse care, infusion material (immunoglobulin costs were not included as they are fixed by social insurance and identical for all routes) and transportation were the main variables identified as having an important impact on costs difference. Their relative importance is presented on a tornado diagram, showing that material is the first cost driver (Fig. 1). Indeed, the number of infusion pumps used has a tremendous impact on cost difference.
Table 1. Baseline case yearly costs.
NA: not applicable; IVIG: intravenous immunoglobulin; SCIG: subcutaneous Ig infusions.
Total direct costs
Table 2. Parameters for sensitivity analysis (yearly costs).
Base case parameters (€)
Sensitivity analysis range (€)
Hospital and transportation costs depend upon the periodicity of infusions, ranging from 21 to 28 days (values in parenthesis are given for a period of 28 days).
VSL (Véhicule Sanitaire Léger) are non-medicalized vehicles used for patients' transportation.
AMI (Acte Médico-Infirmier) are payments units for liberal nurses defined by the French social insurance.
Pumps are rented and the French social insurance has fixed the amount of reimbursement at €306·41/month per unit. IVIG: intravenous immunoglobulin; SCIG: subcutaneous infusions.
Thirty-seven patients answered the questionnaire. One patient was not included in the study because he was switching from IVIG to SCIG, hence introducing confusion. Patients' general characteristics are displayed in Table 3. Seventy-two per cent (26 of 36) of patients received IVIG in an out-patient setting, 22% (eight of 36) SCIG at home and 6% (two of 36) had their i.v. infusion home-delivered. Mean durations of current treatment were 10·1 and 2·5 years for IVIG hospital based and SCIG, respectively. Minimum duration was found for a patient receiving SCIG with 1·2 years. From this point, statistical tests were performed only to compare hospital-based IVIG and SCIG. We considered that there were too few patients receiving IVIG at home to draw conclusions from their analysis. The distribution of diseases (HIGM and agammaglobulinaemia) was identical for both routes of administration. Median ages and weight were also not statistically different. Although immunoglobulin trough levels were similar in i.v.- and s.c.-treated patients, the dose required to produce that level was significantly higher in i.v.-treated patients. Consequently, mean immunoglobulin cost, total mean direct cost and total mean cost were also significantly higher for IVIG. Mean immunoglobulin cost represented between 65 and 90% of total mean cost, depending upon the route of administration. All patients with SCIG used two infusion pumps (eight of eight) but none of them asked for the help of a nurse. Conversely, both patients receiving IVIG at home needed nurse care. Thirty-four per cent (nine of 26) of patients with hospital IVIG were said to be covered incompletely by social insurance. In these cases, transportation was left at the patient's charge. Seventeen of 33 had experienced another route of administration or another preparation. While all patients with SCIG have previously tried IVIG (eight of eight), only two patients with IVIG have experienced SCIG. The former pointed out autonomy and rapidity as the main causes for switching to SCIG. The latter complained about the frequency of s.c. infusions and local reaction. There was no statistical difference for satisfaction scores between the two groups (Table 4). Both agreed for a negligible impact of side effects, but also complained similarly about convenience.
Table 3. Patients' data.
P < 0·05 when compared with subcutaneous infusions (SCIG). HIGM: hyper-immunoglobulin (Ig)M syndromes; IVIG: intravenous immunoglobulin; NA: not applicable.
Our results are consistent with previous studies, with mean cost of immunoglobulin accounting for nearly 75% of direct medical cost. Because monthly doses were assumed to be equal for both routes of administration in the simulation, it had no impact on cost differences. Moreover, this simulation shows that SCIG and IVIG (hospital-based) costs are very similar. It also suggests that home-based IVIG could be the least expensive modality. One-way sensitivity analyses underline the weight of infusion material through the number of pumps used. Pumps are rented monthly and the rent is directly proportional to the number of pumps. This result underlines the importance of local studies, as these costs differ highly from one country to another. The decrease of infusion time is proportional to the number of pumps, increasing patient utility. A trade-off has to be made, and it could be imagined that social insurance would define a fixed number of pumps that would be reimbursed. Little can be done to reduce transportation costs except to improve the recent development of regional centres able to deal with these patients. Distance to competent hospital departments should not exceed 50 km. When clinically possible and tolerated, a 28-day periodicity for IVIG could be proposed, improving the patient's quality of life and reducing total costs. Therefore, as long as the patient is autonomous and no extra pump is prescribed, the choice of the route of administration could be made without any economic barrier.
Calculations made with real data raise interesting issues. First, and probably the most important, the assumption that both routes shared the same dose of immunoglobulin was proved to be inaccurate. Mean immunoglobulin dose was significantly higher (50% higher) for hospital-based IVIG compared to SCIG. Home-based IVIG mean dose was found to be even twice as high as SCIG, but relying merely upon two patients with higher mean weight. Statistical significance was reached for absolute monthly dose, which drives the cost, but not for monthly dose per weight unit (645 versus 817 mg/kg per month, P = 0·16, for SCIG and IVIG hospital-based, respectively). Nevertheless, this finding remains remarkable, as a previous study suggested that the s.c. dose should be 137% of the previous i.v. dose . A supposed superiority of the i.v. route in terms of bioavailability was then pointed out. As mentioned previously, patients with SCIG were not different in terms of disease, age, weight and IgG trough level. However, it is not possible to be fully conclusive regarding the severity of their deficiency. It could be argued that patients receiving IVIG suffered from a more significant antibody deficiency before initiation of immunoglobulin replacement but this was not the case, based on pretreatment IgG serum levels. In any case, doses prescribed in our patients appear to be high with regard to previous studies, which considered 600 mg/kg per month to be a high dose [4,20].
Secondly, nurse care and transportation costs appeared to be negligible. The two patients having nurses provide home visits for IVIG treatment appear to be counter-examples, as home-based treatment is supposed to improve the patient's autonomy. Lastly, as predicted, infusion pumps represented an important share of SCIG costs (one-third). To conclude, results of the simulation were found to be very accurate for hospital-based IVIG, while SCIG costs were overestimated because the immunoglobulin dose was found to be lower than expected. Further and more specific studies are needed to confirm these findings. Out-of-pocket payments only concerned transportation costs. Because all patients had comprehensive coverage, this result was not expected. This problem should be addressed even if it probably carries little impact upon care.
Unsurprisingly, similar global satisfaction was found for both routes of administration. Immunoglobulin replacement changed the lives of these patients radically. Patients are satisfied with their treatment, but still complained about convenience. The reasons may differ between SCIG and IVIG, but could not be identified with the TSQM questionnaire. TSQM was chosen for its brevity and simplicity, but appeared to be confusing for many patients (use of double negatives). Therefore, these results should be interpreted carefully. It could be interesting to investigate patients' satisfaction with other modalities, such as SCIG every 2 weeks  or IVIG every 28 days.
We are aware that the main weakness of this study is the limited size of the population, which concerns mainly teenagers and young adults. The burden for informal caregivers was also not taken into account. Our observation set the basis for a larger-scale study including patients suffering from PID diagnosed during adulthood, such as CVID.
While the theoretical model showed very little difference between SCIG and hospital-based IVIG costs, according to the field data SCIG appears to be 25% less expensive. This is due mainly to an important difference in terms of doses of immunoglobulin. The reality of this difference for larger cohorts of patients needs to be confirmed by further studies. Some attention should also be given to the number of pumps used for home-based infusions. This would not necessarily be the case in other countries, where the supply and cost of pumps differ from France. Findings provided by the theoretical model can be applied reasonably to other primary antibody deficiencies such as CVID, as doses of immunoglobulin were assumed to be equal for all routes of administration. However, patients with CVID may require higher doses of immunoglobulin because of frequently associated chronic lung diseases or enteropathy . Overall costs may therefore be higher.
This study was supported by a grant from the French Ministry of Health. It was also supported by the French PID patients' association IRIS. The CEREDIH received support from LFB, Baxter, CSL Behring and Octapharma.
None of the authors has any conflict of interest with the subject matter or materials discussed in the manuscript.